SOLIDIFICATION OF IRON BASE ALLOYS of iron base alloys at large degrees of undercooling interim report ammrc cr 69-14/1 by w. e. brower, jr. and m. c. flemings forAuthors: Hasse Fredriksson

-··· -·-··~----~ ·-:--:--.. __ .,..... ~,c-=::.· ·-·- ···-··=-.....,-....,...,_...-_________ _

SOLIDIFICATION OF IRON BASE ALLOYS

AT LARGE DEGREES OF UNDERCOOLING

INTERIM REPORT

AMMRC CR 69-14/1

by

W. E. Brower, Jr. and M. C. Flemings

for

Contract Period

January, 1968 - January, 1969

Department of Metallurgy Massachusetts Institute of Technology

Cambridg~, Massachusetts 02139

Julv 15, 1969

DA-46-68-C-0044

~;

D/A Project No. 1C024401A328 AMCMS Code No. 5025.11.294

Metals Research for Army Materiel

This document has been approved for public release and sale; its distribution is unlimited.

u. S. Army Materials and Mechanics Center Watertown, Massachusetts 02172

SOLIDIFICATION OF IRON BASE ALLOYS

AT LARGE DEGREES OF UNDERCOOLING

INTERIM REPORT

AMMRC CR 69-14/1

by

W. E. Brower, Jr. and M. C. Flemings

for

Contract Period

January, 1958 - January, 1969

Department of Metallurgy Massachusetts Institute of Technology

Cambridge, Massachusetts 02139

Juiv 15, 1969

DA-46-68-C-0044

D/A Project No. 1C024401A328 AMCMS Code No. 5025.11.294

Metals Research for Army Materiel

This document has been approved for public release and sale; its distribution is unlimited.

U. S. Army Materials and Mechanics Center Watertown, Massachusetts 02172

rz ..7«. - .

i.

ABSTRACT

A levitatlon melting and splat cooling apparatus has been modified to permit chill casting and liquid quenching. It was utilized in this work to investigate the effect of cooling rate on solidification structure and mechanical behavior of iron base alloys.

The degree of microstructural refinement of several alloys was dttennined quantitatively and qualitatively over a range of cooling rates from l8C/second to 10^0C/second. Alloys studied Include Fe-25% Ni, 440C, 4330, Fe-4.3%C, Fe containing SIO2 inclusions and Fe containing FeO inclusions.

SIO2 inclusions in Fe alloy were observed to be measurably coarsened after isothermal solid state heat treatment at 14000C.

The fracture behavior of inclusion bearing pure iron was investigated using the scanning electron microscope. Studies were on two alloys, Fe containing SiC^ inclusions, and Fe containing FeO Inclusions. In both alloys, the ductile fracture surface revealed dimples containing spherical inclusions. Extreme Inclusion-matrix interface separation giving a "ball in a trough" appearance resulted from a rotational deformation mode during necking down of the test bar. Dimples immediately associated with the inclusions were smaller the smaller the inclusion. Fracture surfaces of the two alloys were similar. Microcracks associated with FeO inclusions in the Fe-0.07% 0 alloy were observed in the as-cast structure, while no microcracks associated with the SIO2 inclusions in the Fe-0.05% Si alloy were observed in the as-cast structure.

FeO inclusions exhibit shrinkage cavities. These result because the Inclusions form as a liquid and solidify after the iron. Inclusion- matrix separation of FeO inclusions Is also observed because of contraction of the FeO inclusions on cooling. Shrinkage cavities or Inclusion-matrix separations were not observed in the alloy containing SIO2 Inclusions (in the as-cast condition). Inclusion-matrix separation was, however, observed for both types of inclusions Jn the plastically deformed region of a test bar after testing.

TABLE OF CONTENTS

ii

Section Number

CHAPTER I.

CHAPTER II.

ABSTRACT

LIST OF ILLUSTRATIONS

LIST OF TABLES INTRODUCTION REFERENCES

VARIATION OF STRUCTURE WITH COOLING RATE

Inti-oductiou and Literature Survey

Apparatus-General

Levitation System

Controlled Atmosphere and Gas Cooling System

Quenching Mechanisms

Cooling Rates During Solidification

Metals and Alloys Studied

Results - Matrix Microstructure, Inclusion Size and Morphology

Discussion

ISOTHERMAL SOLID STATE COARSENING OF SILICA INCLUSIONS

Introduction and Literature Survey

Procedure

Results and Discussion

Page Number

i

iv

vii vili xl

1

1

4

5

8

8

12

14

15

22

51

51

53

54

iii

Section Number

CHAPTER III. M.CHANICAL BEHAVIOR OF INCLUSION BEARING IRON


Procedure

Results

Discussion

CONCLUSIONS

APPENDIX A

APPENDIX B

APPENDIX C

APPENDIX D

Page Number

61

61

64

6!,

68

91

93

9A

96

98

vl

Figure Number

3-10

3-11

3-12

3-13

3-14

3-15

3-16

3-17

3-18

3-19

C-l

Electropollshed surface of the Pe-0.05^31 master alloy

Electropollshed and mechanically polished surfaces of the Pe-0.07^0 master alloy. ...

Electropollshed surface of 0.05fl»Sl-Pe master alloy 1/2 Inch gage length test bar, strained to, 3$ elongation. ..

Electropollshed surface of the Pe-0.05^31 master alloy 1/2 inch gage length test bar, strained to necking. . .

Mechanically polished burface of the Pe-O.OSßSl master alloy 1/2 inch gage length test bar, strained to fracture

Electropollshed surface of Pe-O.OTÔ master alloy 1/2 inch gage length test bar, strained to fracture, .

Electrcoolished surface of Pe-0.05^31 chill casting 1/4 inch gage length test bar, strained to fracture. .

Electropollshed surface of Fe-0.07^0 chill casting 1/4 inch gage length test bar, strained to fracture.

Schematic description of the deformation associated with PeO and 310» inclusions

Thermal expansion of pure iron, FeO, and 310_.

Size distribution conversion for FeO inclusions ....

Page Number

81

82

83

84

85

86

87

88

89

90

98

--■

Figure Page Number Number

1-21 Size distributions of Si0o inclusions 46 2 I

1-22 Mean measured Si02 inclusion diameter versus cooling rate 47

for PB-0.05$Si alloy.

1-23 Variation of microstncture with cooling rate for Pe-Mn-S 48 alloy.

1-24 Variation of inclusion morphology with cooling rate for 49 4330 low alloy steel.

1-25 Variation of microstructure with cooling rate for 50 Pe-25^Ni-0.05^31 alloy.

2-1 Size distributions of Si0o inclusions in the "ê-O.O^Sl 57 master alloy.

2-2 Size distribution of Si02 inclusions in the Pfe-O.05^81 58 chill castings. ,

2-3 Size distributions of Si02 inclusions of the Pe-0.05#Si 59 alloy splats.

2-4 Variation of Inclusion size 60

3-1 Fracture surface of Perrovac "E" 72

3-2 Fracture surface of Pe-0.05#S1 alloy 73

3-3 Fracture surface of Fö-0.05$Si chill casting 74

3-4 Fracture surface of Fe-0.05^Si splat . . 75

3-5 Fracture surface of Pe \ 07210 master alloy 76

3-6 Fracture surface of Pe-0.07/S0 chill casting . 77

3-7 Fracture surface of Pe-0.07^0 splat 78

3-8 Necked region of the fracture surface of the Pe-0.05^Si 79 master alloy

3-9 Necked region of the fractur«? surface of the Fe-0,07^0 80 master alloy

Iv

LIST OP ILLUSTRATIONS

Figure Page Number Number

1-1 Sketch of levitation melting and casting apparatus. 26

1-2 Photograph of levitation melting and casting apparatus. 27

1-3 Photograph of levitation coll. 28

1-4 Schematic diagram of levitation melter circuit. 29

i 1-5 Schematic diagram of temperature measuring system. 30

j 1-6 Schematic diagram of gas flow system. 31

1-7 Photograph of "hammer and anvil". 32

1-8 Optical photocell eye cooling curve 33

1-9 Thermocouple cooling curve for Fe-2,$Ni alloy 34

1-10 Pe-25$Ni chill plate casting 35

1-11 The variation of microstructure with cooling rate for 36 Pe-255i5Nl alloy.

1-12 Variation of microstructure with cooling rate for 440C 37 alloy.

1-13 Variation of microstructure with cooling rate for 4330 38 alloy.

1-14 Dendrite arm spacing versus cooling rate, Fe-25 per cent 39 Nl alloy.

I-I5 Dendrite arm spacing versus cooling rate, 4400 alloy. 40

l-l6 Variation of microstructure with cooling rate for 41 Pe-4.3^C alloy.

1-17 Lamellar spacing versus cooling rate, Fe-4.3^C alloy. 42

l-l8 Microstructure of Fe-0.05S6Si master alloy. 43

1-19 Microstructure of Pe-0,07^0 alloy. 44

1-20 Variation of microstructure with cooling rate for 45 Pe-0.05^Si alloy.

fe

vii

LIST OF TABLES

Table Page Number Number

1 Quantitative Metallography Results. 18

D-l Mechanical Property Data. 100

viii.

INTRODUCTION

This report summarizes one of several related research activities

on inclusions.at Massachusetts Institute of Technology. The work is

primarily on levitation melted and rapidly solidified (including splat-

cooled) samples. It deals with effects of solidification variables on

(1) dendrite and inclusion structures, (2) influence of high temperature

bomogenization treatments on inclusions, and (3) influence of inclusions

on fracture behavior of iron.

The work is an outgrowth of a program on undercooleü iron base

alloys, that has had important technological, as well as fundamental,

findings. A most important fundamental result has been that structure

coarsening ("ripening") determines final dendrite arm spacing in both

undercooled melts and in usual commercial castings and ingots. Dendrite

arm spacing is of major importance in determining properties of castings

and of wrought material produced from cast ingots and so determination

of the mechanism of establishment of dendrite arm spacing has been of

considerable applied as well theoretical interest.

Most recent work on this program, prior to that reported herein,

has been on extension of the dendrite coarsening ideas to inclusion

morphology and growth. This work has shown that inclusions coarsen

(Ostwald ripen) during solidification, just as do dendrites. Inclusions

also are "pushed" by growing dendrites, and float, join and coaJesce during

ix.

(4) solidification. Initial work has l>een reported and continuing work

will be summarized in a later report. Work to date has been primarily

on iron-copper alloys with silica inclusions.

Chapter I deals with effects of cooling rate on dendrlte and

inclusion structure. A levitation melting, undercooling, and splat

cooling device is employed in this work. Design and construction of

the unit are described herein; portions of this description have been

(A) given in the previous annual contract report . Some particularly

interesting results from this work Include the linear relationship,

on log-log scale, extending over many orders of magnitude, between

dendrlte arm spacing and cooling rate (Figures 14 and 15) and between

inclusion diameter and cooling rate (Figure 22).

Chapter II deals with isothermal coarsening of SiO^ inclusions

in nearly pure iron. It is of Interest that appreciable coarsening of

these Inclusions occurs after a relatively few hours at 1400oC, This

is approximately the temperature used in long time heat treatments by

Quigley and Aheam, It therefore seems probable that some of the bene-

ficial effects of these high temperature heat treatments on properties

are due to their influence on inclusions.

Chapter III summarizes work on mechanical behavior of several

inclusion bearing irons. This chapter includes a number of scanning

electron micrographs of fracture surfaces and polished sections, showing

X.

inclusions and relation of inclurions to fracture. Inclusions were

found in a large fraction of the dimples on all fracture surfaces

(e.g.. Figure 2, Chapter III). An Interesting difference between FeO

and SiO- inclusions in as-cast material was shown to be that the FeO

inclusions exhibit solidification shrinkage cavities and tend to pull

away from the matrix during cooling (e.g.. Figure 11, Chapter III).

xi.

REFERENCES

1. Castings and Solidification Section, Department of Metallurgy, M.I.T., "Solidification of Iron Base Alloys at Large Degrees of Undercooling", U. S. Army Materials Research Agency, Contract No. DA-19-020-AMC-0231(X), Interim Report, October 1, 1963- September 30, 1964.

2. T. Z. Kattamis, M. C. Flemings, "Solidification of Iron Base Alloys at Large Degrees of Undercooling", U. S. Army Materials Research Agency, Contract No. DA-19-02o-AMC-0231(X), Interim Report, October 1, 1964-September 30, 1965.

3. T. Z. Kattamis, M. C. Flemings, "Solidification of Iron Base Alloys at Large Degrees of Undercooling", U. S. Army Materials Research Agency, Contract No. DA-19-020-AMC-0231(X), Interim Report, October 1, 1965- September 30, 1966.

4. M. C. Flemings, M. Myers, and W. E. Brower, Jr., "Solidification of Iron Base Alloys at Large Degrees of Undercooling", U.S. Army Materials and Mechanics Research Center, Contract No. DA-19-020-AMC-0231(X), Interim Report, October 1, 1966-Decemher 31, 1967.

5. T. Z. Kattamis, M. C. Flemings, "Dendrite Structure and Grain Size of Undercooled Melts", Trans. Met. Soc. AIME, November, 1966, pp. 1523- 1532.

6. T. Z. Kattamis, M. C. Flemings, "Solidification of Highly Undercooled Castings", Trans. AFS, Vol. 75, 1967, pp. 191-198.

7. T. Z. Kattamis, M. C. Flemings, "Structure of Undercooled Eutectics", to be published.

Chapter I

VARIATION OF STRUCTURE WITH COOLING RATE


Interest In higher cooling rates during solidification has led

In recent years to the development of a quenching technique known as splat

cooling, reviewed In Duwez's Cambell Memorial Lecture . Duwez and co-

workers have investigated many alloy systems by the "shock tube" technique Q 5 i| C £ •T

of splat cooling . Predeckl has measured and Ruhl has calculated

the cooling rates during splat cooling by the shock tube method to be

10-10 C/sec. Non-equilibrium structures that result from this technique

are either non-equilibrium single phase alloys, i.e., extended solid

2 3 solubility , single phase eutectic alloys , or new non-equilibrium phases

unobtainable for the given alloy by slower quenching techniques, i.e.,

4 5 8 new intermetallic compounds and amorphous metal alloys . Ruhl and Cohen

have produced a new phase In the iron-carbon system using the shock tube

technique. The splat resulting from this technique is a layer of small

flakes approximately one micron thick. In general, this type of splat,

although extremely well suited to X-ray and electron diffraction analysis,

is not suitable for optical microscopy or mechanical testing in the as cast

condition.

The other predominant splat cooling technique, used in this work,

is the "hammer and anvil" technique which results in approximately a one

inch diameter by 100 micron thick disc. Gtrachan has measured the cooling

rate during splat cooling in the hammer and anvil device used In this work

to be on the order of 10 C/sec. . The hammer and anvil yields a splat

that Is 100 times slower cooled but 100 times larger than wie product of

the shock tube. Both optical microscopy and ..lechanlcal property measure-

ments can be performed on the hammer and anvil type splat. Some examples

of non-equilibrium structure? that have been obtained from the hammer and

anvil splat cooler are: amorphous Fe-C-P alloys , single phase Ni-32£Sn

eutectlc alloy and extension of the equilibrium solid soludility of

12 copper in silver . Ruhl and Cohen's epsllon phase in the iron-carbon

system, however, was unobtainable on the hammer and anvil type splat cooler

used in this work.

The main goal of the splat cooling technique employed in this

research was the achievement of ultra-fine dendrite structures and

inclusions as compared with slowly solldlfiec' structures. Refinement of

dendrite structure has been shown, for several alloys, to be solely a 13

function of cooling rate (or "local solidification :ime"), and hence the

only way to achieve ultra-fine structures is, apparently, by very fast

cooling.

Fine dendrite arm spaclngs In cast alloys are desirable, since

it has been shown that the homogenlzatlon time for an alloy with non-

equilibrium solute segregation is proportional to the square of the

:ca:

.15.16,17

lb secondary dendrite arm spacing . It is also known that the mechanical

properties oi' a heat treated casting depend on the homogeniety obtained

Rirthermore, as cast segregation Is present in wrought material and is

deleterious to properties ' y , Dendrite arm spaclngs ranging from 50

microns down to less than 1 micron have been achieved with the increased

cooling rates available in the present apparatus. Segregation in the

form of non-metallic Inclusions Is also refined by higher cooling rates.

To establish continuous trends of structural variation with

cooling rate, solidification techniques with cooling rates intermediate

between conventionally solidified, slowly cooled large castings and ultra

rapidly cooled splats were utilized, "nils necessitated a redesigning of

the existing levitation melting and splat cooling unit. Small thin plate

chill castings, although extremely finely structured, have much coarser

structures and have over an order of magnitude slower cooling rate than

the splats. Chill casting capability was therefore Incorporated into the

levitation melting find splat cooling apparatus. Still coarser structures

and slower cooling rates result from liquid quenching of the molten

levitated charge. The ability to replace the splat cooler with a liquid

quench tank in the redesigned apparatus yields a quenching technique with

a cooling rate intermediate between that of chill casting and conventional

slow solidification. With these four quenching techniques available-

conventional casting, liquid quenching, chill casting, and splat cooling—

continual refinement of solidification structures may be studied over six

orders of magnitude of cooling rate, 1 C/sec. to 10 C/sec.

Apparatus-General

Extensive work was done to modify and extend the capabilities

of an existing levitatlon melting unit. This work was primarily to

(1) Improve the vacuum system and purity of cooling gases employed,

(2) Incorporate ability to make chill plate castings and to liquid quench

the molten droplets, and (3) permit making up to five separate runs

without breaking vacuum. The capabilities of the apparatus Include the

following:

1. Ferrous metals or alloys can be levitatlon melted in amounts

up to about 3 grams In vacuum or controlled atmosphere. A wide variety of

other metals can also be so melted.

2. Samples can be held molten or partially molten for long

periods (e.g., in excess of l/2 hour). Large superheats can be obtained

(in excess of 600 C). Large undercoolings can be achieved (as much as

3000C under the melting point).

3. Alloying can be accomplished by simultaneously melting two

or more charge materials (while levitated) or by reaction of the charge

with t controlled atmosphere. Alternately, the charge can be pre-alloyed

in the vacuum induction furnace.

4. Up to five separate melting and casting runs can be made

without opening the apparatus to tha atmosphere.

5. The metal charge can be solidified by (a) slow controlled

cooling in the levitatlon coil, (b) liquid quenching, (c) casting in thlri

section ingot molds, and (d) "splat cooling".

6. Interrupted solidification experiments can be conducted by,

for example, partially solidifying the sample (slowly) In the coll and

then rapidly solidifying the remainder by liquid quenching or splat

cooling.

The basic levltatlon melter employed has been previously

20 21 described ; It is based on the work of Comenetz and others . The

Q "hammer and anvil" splat cooling device has also been previously described .

The propulsion of the hammer is electromagnetic. A large current is

released for a short time through a "pancake" coil adjacent to a driving

disc. The driving disc Is then repulsed from the coll with a force that

may be calculated from magnetodynamlc principles. In the splatter the

forct achievable is 2200 to 6600 pounds; resulting platen velocity is

1000 to 3000 cm/sec. (33 to 100 ft/sec). Remaining portions of the

apparatus have been constructed in the course of this research and are

described in the following sections.

Figure 1 Is an overall sketch of the invitation melting and

quenching apparatus, and Figure 2 (top) a photograph of the apparatus,

including power supply. Descriptions are given below of the melting,

temperature measuring, atmosphere controlling, gas quenching, liquid

quenching, chill casting, and splat quenching Systems». Figures 3-7

show details of some of these systems.

Levltatlon System

The levltatlon melter constructed for this work was designed to

be operated with a 10 KW, kOO K.C. Lepel High Frequency Generator. This

power source is capable of an output of approximately 300 amperes into a

suitable Inôdancc load, but is limited to less than this Into the relatively

low inductance levltatlon coll. Figure 3 shows a levltation coil of the

type presently in use; it is made from 1/8 inch diameter, thin wall copper

tubing.

The most practical technique for temperature control during

levitation involves flowing a stream of cooling gas over the levitated

charge. This is readily done by winding a levitation coll, such as the

one in Figure 3, around a vycor tube of large enough diameter to contain

the specimen, and then flowing cooling gas vertically up the tube.

To obtain good "matching" to the generator and to Increase the

circulating current in the coil, a capacitor bank (Lepel CT-25-4) was

Inserted Into the output circuit in parallel with the levitation coil.

The coll and the capacitor bank form a resonant circuit, having a large

circulating current, measured to be over 650 amperes at full power.

Physically, the capacitor is located close to the levitation coil and

they are connected using co-axlal power leads to minimize power losses.

A schematic diagram of the external electrical circuit is given in Figure

•Die enclosure containing charges and ingot molds is shown

schematically In Figure 1, This enclosure is a duplicate of that designed

27 by Yarwood except modified for splatting, as described below. Up to five

charges arc placed in each charge container prior to evacuating the system.

The charge container is of boron nitride to prevent Its being heated when

raised into the levitation coil. Charging is accomplished by first

rotating the turntable into position so one of the five charges is

directly beneath the vertical glass tube. Next, the charge and charge

container are raised up into the levitation coil by pushing on a vertical

pushrod which extends out the bottom of the box. This pushrod also serves

as a "charge exit pDrt'1 for splatting and is so named in the sketch. An

"O" rinc; coal betwtitrn the pushrod and enclosure allows the lateral motion

«Mie iwafntalnlng vacuun or slight positive pressure In the enclosure.

turatahle also coBUdüas up to five ingot aolds. lb cast

of these, it is shapij rotated under the coll wWle the droplet is

levitated. The pcwer is tamed off and aetal dropped into the plate,

lb "splat* a snfile, one of the locations for an ingot nold is left

vacant, and this location placed under the levitated drop. Wien the

levitation pouer is turned off, the droplet passes thronî the hole, doan

the hoi Iff* "pushrod-charBe exit port', and through * plastic sheet sealed

at the base of the exit port. This sheet plastic seal covers a 5/B"

hole; ve— or positive pressure is again —intained by an *0^ ring

seal. Hie seal is "Saran Mrap" (polrvingrlidene chloride copoljaerlaed

with poljvinfl chloride); it is sufficiently strong to resist ataospfaeric

pressure ahile the enclosure is fblly evacuated, but rapidly nelts as the

falling hot drop appruadics it.

The teaperature of the levitated droplet is uonitored using a

IHlletran Tbo-Cblor l^frmeter. Model TSA. Its accuracy varies with the

aaterlal whose tenperature is being neasured. but has been fomd in this

(and with this systcu) to be accurate on absolute ueastwrwaents on

Uqnld Iroaa and steel to within + 20OC and to within + 10oC relative to

the neasured nelting point. Output is read on a neter in the control

unit, and nay be recorded on an external chart recorder. A sebeuatic

diagrau of the tenperature ueasuring systen is shown in Figure 3«

Sighting of the optical pyroneter is done from the top of the levitation

coil through a right angle prlaa and a flat glass disc glued to the top

of the glass tube surrounding the specinen. The disc is well above the

levitation coll (3 inches) to prevent deposition on it of vapors fTna the

levitated cnat-Ge.

8

Controlled Atmosphere and Gas Cooling System

A schematic diagram of the controlled atmosphere and gas

cooling system is shown in Figure 6. Evacuation of the system is

accomplished by means of mechanical and diffusloi. pumps, acting through

a 1-3/8 inch hole in the bottom of the enclosure containing the charges

and ingot molds. A thermocouple vacuum gauge is permanently mounted on

the enclosure« and a vacuum ion gauge may be attached to one of the two

standard vacuum fittings.

The desired gas atmosphere (or combination of gases) Is

admitted through copper tubing to flow rate control and solenoid-

operated check valves. Ability to admit a large and variable flow rate

(up to 250 cubic feet per hour) to the system permits temperature control

of the levitated charge. The cooling gas passes up the 5/8 inch I.D.

Vycor tube surrounding the levitated specimen and exits through a vacuum

released valve. This valve acts as a vacuum seal when the system is

evacuated. All connections in the gas flow system are 3/8 inch copper

tubing to prevent gas contamination.

Quenching Mechanisms

The various ways summarized below of solidifying the samples

permit obtaining cooling rates of the orders of: 1 - 10oC/sec. for gas

o ^ 4 quenching} 30 - 200 C/sec. for liquid quenching, 10J - 10 oc/sec, for chill

5 6 casting, 10 - 10 oc/cec, for splat cooling.

With sufficiently high flow rates of hydrogen or helium it is

possible to solidify a molten levitated charge. At the relatively slow

rates of cooling normally Involved, the cooling rate may be measured from

the chart recorder of the optical pyrometer. This cooling rate simulates

those observed In relatively large castings. As noted previously,

quenching into liquid is achieved by placing a liquid quench tank under

the charge exit port in place of the splat cooler, and dropping the

charge through the plastic seal.

To achieve still higher cooling rates, copper molds with plate-

shaped mold cavities of varying thickness are inserted in the turntable

in the enclosure in Figure 1 . These copper molds are one inch diameter

split cylinders with a wedge or conical shaped riser section above the

plate cavity and 1/64 inch vent holes below the plate cavity to aid in

mold filling. Plate thicknesses are .03 inches, .05 inches, and .08

inches. Por best mold filling, molds are polished eaih time and minimum

superheat of 100 C employed.

An important feature of the arrangement for "splatting" shown

in Figure 1 is that the rather delicate levitation melting device and

its attachments are physically separate from the necessarily violent

clapper (clapping velocities are up to 200 miles per hour).

The drive mechanism for the "hammer and anvil" type splat

cooler shown in Figure 7 consists of a driving coil and capacitor energy

storage bank, a power supply to charge the capacitor bank, droplet sensing

equipment, and spark gap switch with associated trigger circuitry. The

capacitor bank and associated power supply are manufactured by EG & G

International, Inc. The capacitor bank consists of two separate but

equivalent units. Models 524 and 233» each containing capacitors totalling

320 microfarads and having a storage capacity of 2500 Joules at 4,0 KV.

These two units are connected in parallel, and to the power supply.

Model 522, which has a 0 - 4,0 KV adjustable voltage output. The power

10

supply unit also contains circuitry to provide a voltage pulse output

to trigger the spark gap, upon an external contact closure.

The droplet sensor Is a Knight photoelectronlc relay (kit).

A CdSe photocell is located close to the clapper platens. In the line

of sight of fall of the molten droplet. When activated by a falling

droplet, a relay contact Is closed, and through connection with the power

supply trigger circuit, provides a voltage pulse to trigger the spark gap.

The spark gap unit consists of an EG & G experimental "rail-type" spark

gap and an EG & G trigger transformer. The pulse from the power supply

is transformed to a much higher voltage, which ionizes the spark gap and

causes electrical breakdown, rendering the spark gap conducting, and

discharging the capacitor bank into the driving coll.

Upon sensing a falling droplet, the spark gap is triggered and

the energy stored in the capacitor bank is released into the driving coil.

The large current pulse through the driving coil, associated with the

capacitor discharge, creates an intense magnetic field around the coil

which in turn Induces eddy currents in the aluminum driving disc. The

eddy currents cause another magnetic field out of phase with the first.

Thus, the driving plate is repelled from the coil at high speed. The

capacitor discharge primary pulse lasts approximately 0.1 - 0.2 milliseconds;

the moving elements ire accelerated to maximum speed during this time.

Operation of the splat cooler and associated equipment is

briefly as follows. The clapper platens are cleaned and the moving platen

assembly is set in the open position with driving plate against the driving

coil. The droplet soncins photocell is Inserted into position in the side

of the protective cover and the cover is placed over the clapper. After

11

the induction power is turned on and the specimen levitated, the clapper

power supply is energized to charge the capacitor bank to the desired

voltage. Specimen temperature is manipulated by changing gases or gas

flow rate. When proper conditions have been attained, the induction power

is switched off. The specimen falls and is sensed by the photocell which

triggers the capacitor discharge. The moving platen smashes the falling

specimen against the back platen to form a splat. Total elapsed time

between sensi;ig the specimen by the photocell and splattlng is about five

milliseconds.

12

Cooling Rates During Solidification

As mentioned In the literature survey, Strachan has measured

the cooling rate obtained in the splat cooler used In this Investigation

by use of a photovoltaic cell. A typical cooling curve Is shown In

Figure 8, yielding a coolli« rate from 1500OC - 13000C of 10 0C/sec.

To measure cooling rates during solidification of the plate

castings, a small Pt-PtlO^RH thermocouple (Number 38 thermocouple wire)

embedded In the side of the plate-shaped cavity of the copper chill molds

was used. The output of the thermocouple was recorded on an oscilloscope,

giving a cooling curve similar to those obtained in splat cooling. Figure 9.

The measured rate from 1500OC - 13000C of 1300OC/sec. of a O.O80 inch

thick chill plate casting is necessarily an average value, since as seen

by variation of the microstructure of the resulting Fe-25^Ni casting,

shown in Figure 10, the cooling rate (as characterized by tha secondary

dendrite arm spacing) varies over the thickness of the chill casting.

The value of the cooling rate for the type of quenching oil used

in this investigation is calculated in Appendix B to be 140 C/sec. As

heat, cransfer from the molten metal drop to the liquid (through a vapor

boundary layer) is "h-eontrolled". Appendix B, the cooling rate is uniform

throughout the drop.

During sac quenching in the levitation furnace the output of the

optical pyrometer may be displayed on a strip chart recorder. Cooling rates

of 1 - 10OC/sec. may be obtained with a sufficiently high (250 SCFH) flow

rate of hydrogen or helium.

13

During solidification In a crucible in the vacuum induction

furnace, the output of a Pt-PtlQURh thermocouple immersed In the melt

is displayed on a strip chart recorder. Cooling rates of 0.1 - 1.5 C/sec,

are possible.

—.r... .o.vnem**'

14

Metals and Alloys Studied

The following materials were either studied as received or used

as alloying material:

1. Perrovac ME" Iron, whose chemical analysis as provided by

Its producer, the Crucible Steel Company, Is given in Appendix A, was

used as the principal alloying element and was mechanically tested in

tension in the as received and completely annealed conditions.

2. Electrolytic nickel, whose chemical analysis is given in

Appendix A, was used in the high purity Pe-25S&11 alloy.

3. "Hie following elements and compounds were used in reagent

grade purity: Silicon, ferric oxide, carbon, copper, sulfur, arvl

manganese.

4. The chemical analysis of commercial 440C alloy is given in

Appendix A.

5. The chemical analysis of commercial 4330 alloy Is given in

Appendix A.

The following alloys were prepared from the above prime materials

in the vacuum induction furnace for use in the levltation furnace:

Pe-4.3wt.^C, Pe-0.05wt.56si, Fe-0.07wt.$60, Fe-Mn-S, Pe-2S8Ni-0.05wt.$Si,

Fe-2595N1. The chemical analyses are given in Appendix A.

In addition, hlch purity Fe-25^N1 alloy was prepared from

previously zone-refintd Perrovac "E" iron and electrolytic nickel rods by

a "zona alloying" technique In the electron beam furnace. The chemical

analysis In given in Appendix A,

15

RESULTS

Matrix Mlcrostructure, Inclusion Size and Morphology

Several alloys were chosen to demonstrate the relationship of

dendrite arm spacing to cooling rate for iron base alloys over a range

o , of cooling rates from that of normal casting of about, i C/sec. up to that

of splat cooling, 10 0C/sec. ("hammer & anvil" technique).

Single phase Fe-25^Ni and multiphase commercial 440C and 4330

alloys were solidified by the four quenching techniques described above.

Photomicrographs of the resulting microstructures are shown in Figures 11,

12, and 13. The dendritic solidification morphology of 4330 alloy is

shown qualitatively to be refined by higher cooling rates. Plots of the

secondary dendrite arm spacing vs. cooling rate are shown in Figures 14 and

22 15 for Fe-2555N1 and 440C alloy. Prom the linearity of the log-log-plots,

one can see that the secondary dendrite arm spacing is a function only

of cooling rate for the alloys studied over six orders of magnitude of

cooling rate. Equations o:p the experimentally determined relationships

are as follows:

Fbr Fe-25^N1. d = 60T"0,32 ^

Fbr 4400, d . 82T"0,41 (2)

* where d is the secondary dendrite arm spacing in microns and T is the

cooling rate in C/sec.

The variation with cooling rate of the structure of a normally

lamellar Fe-4.3^C eutectic alloy (Ledeburite) was also investigated. Figure 16,

16

A relation of the lamellar spacing to the cooling rate similar to the

dendritic oaser was found. The equation of this relationship, shown in

Figure 17 Is as follows:

1 - 2.3T 0-12 (3)

where T is the mean lamellar spacing in microns.

Two types of inclusions in iron were investigated over the range

of cooling rates available. An Ffe-O.05^31 alloy (Perrovac "E" plus reagent

grade silicon) and an Pe-0.075ft) alloy (Perrovac "E" plus re.\gent grade

PBpO.) were prepared in a silica crucible in the vacuum Induction furnace.

Both alloys were allowed to come to equilibrium with the silica crucible

at 1550OC for 20 minutes, then solidified at 1.50C/sec. by turning off the

furnace power. Nucleation occurred after less than 20oC undercooling. The

resulting mlcrostructure of the Pe-0.05^31 alloy, shown in Figure 18

contained, as evidenced by their shape and appearance in white light and

the appearance of the crossed Nicols effect in polarized light, Si0? spherical

glassy inclusions. The resulting mlcrostructure of the Pe-0 alloy, shown

in Figure 19, contained opaque, grey spherical FeO Inclusions. The

equilibrium diagram of the Fe-Si-0 system, and the experimental observations

26 of Forward indicate that the structures obtained are those to be expected

for the two alloys. Pieces of these alloys were then remelted in the

levitatlon furnace and subsequently liquid quenched, chill cast, and splat

cooled. The resulting microstructures are shown in Figures 19 and 20.

In the Fe-0.07$0 master alloy and the oil quenched specimen

spherical FeO inclusions were optically resolvable. Figure 19a and b.

The Fe-0.07?50 chill casting. Figure 19c, shows a distribution of spherical

inclusions not completely resolvable at 1000X magnification. The

17

Fe-O.075^0 splat. Figure 19d, showed no optically resolvable inclusions.

The Fe-0.05^31 master alloy and oil quenched specimen showed spherical

glassy SiOp Inclusions easily identifiable (Nicols effect) in the optical

microscope. Figure 20a and b, Sime the inclusion sizes in the Fe-0.05^31

chill cast specimens and splats were below optical resolution, electron

microscopy of parlodlon replicas of electropolished sections of the

chill casting and splat were used to obtain the microstructures shown

in Figure 20b and c. Although some distribution of inclusion like

particles did appear at 100,OOOX magnification in the splat, their small

size, about 25 angstroms, and their dissimiliarity to those observed in

the chill casting at ll,O0OX magnification makes their identification as

SiOp inclusions difficult. Therefore, no structural measurements were

attempted on the Fe-0.05^31 splat.

The structural parameters measured by quantitative metallographic

analysis were: the mean inclusion diameter on a polished section, d;

the number of inclusions per unit area, NÄ; and the volume fraction of

inclusions, V . For spherical inclusion morphology all other parameters

of Interest may be calculated from these three. Table 1 summarizes the

quantative metallographic results for the as cast specimens discussed

here and the coarsened specimens, discussed in a later section.

To measure d, a size distribution of about 300 measurements of

particle diameters on optical and electron photomicrographs of a polished

section were made and plotted as shown In Figure 21 for the chill cast,

oil quenched, and vacuum Induction furnace melted Fe-0.05^31 master alloy.

The magnification way adjusted so that the mean measured diameter would

be about Itnm. The difference between the size distribution of spherical

18

TABLE 1

QUANTITATIVE METALLOGRAPHY RESULTS

SPECIMEN

NUMBER OP INCLUSIONS MEASURED MAGNIFICATION

HEAD DIAMETER (MICRONS)

SlOg Master , 264 530 2.78

As Cast

Oil Quench, As Cast

331 1,300 0.528

Chill Casting, 393 As Cast

10,000 0.099

Master, 12 Hours

255 330 3.71

Chill, 12 Hours

226 1.230 0.646

Splat, 12 Hours

413 1.090 0.75

Master, 24 Hours

282 325 4.25

Chill, 24 Hours

367 890 1.04

Splat, 24 Hours

238 1.230 0.755

Master, 48 Hours

282 330 4.05

Chill, 48 Hours

245 1,150 0.885

Splat, 48 Hours

388 880 0.98

PeO Master, As Cast

315 250 4.37

NUMBER VOLUME PER UNIT FRACTION AREA 06)

1.3 x 10 0.17

4 x 10° 0.214

2.8 x 10° O.O69

1.2 x 10 0.201

1.67 x 105

1.84 x 105

1.21 x 10 0.210

1.58 x 105

1.97 x 10"

1 45 x 10 0.216

1.48 x 105

1.48 x 105

1.47 x 10 0.174

19

particles presented on a mechanically polished section and an electro-

polished section will be neglected here.

2^ Bergh 4 Lindberg J have presented a method for conversion of

size distribution of spherical particles determined on a plane polish

to a true three-dimensional size distribution. Appendix C is a demonstra-

tion of the technique on the size distribution of the Pe-0.07560 master

alloy. The mean diameter, d, on a polished section is raised from 4.4u

to 5.2w for the true size distribution. Every measured planar size

distribution will be shifted to higher values of d.. Figure C-l, in the

true size distribution, and a factor of 1,2 Increase in mean diameter

will be assumed to be constant for all specimens. Since the relationship

between d and cooling rate, T, is presented as log d vs. log T over six

orders of magnitude, errors of even 20$ in d are not significant.

A plot of log d vs. log T for the Fe-0.05^31 alloy is shown

in Figure 22. The resulting equation is;

d = 4.0T"0,50 (4)

Increased cooling rate also affected the volume fraction and

number per unit volume, N , of inclusions in the Fe-0.05^31 alloy. The

number of inclusions per unit area, N. (directly proportional to the

number per unit volume, N ), was measured directly from the micrographs.

Measurement of volume percent of inclusions was done by a two-dimensional

24 systematic point count as proposed by Hilliard and Cahn . The volume

fraction of Si02 inclusions in the Fe-0.05^31 master alloy and oil quenched

specimen are similar. Table 1, while in the chill casting it is significantly

decreased. Although the volume fraction of 3i0p in the splat was not

measured, it is presumed to be still further decreased from the equilibrium

20

value observed In the master alloy and liquid quenched specimen. With

sufficiently high cooling rate, then the equilibrium as cast volume

fraction of inclusions can be surpressed. The N of S102 inclusions in

the oil quenched specimen of the Pe~0.05^31 alloy is Increased from

the value observed in the master alloy. Table 1, Thcji^pX'IW.Sli V of

SiOp in the chill casting prevents a further increase in N over the

oil quenched specimen. Thus the interparticle spacing (inversely

proportional to N ) is at first decreased, then increased with increasing

cooling rate, due to the combined effects of decreasing inclusion size

and volume fraction.

To demonstrate qualitatively the effect of cooling rate on

other types of inclusions, an Pe-2.056^-1.2^3 alloy and a 4330 low alloy

commercial steel were solidified by the four quenching techniques available.

The concentrations of Mn and 3 in the Pe-Mn-S alloy were chosen so that

only stolchlometrlc MnS Inclusions would precipitate upon solidification.

The microstructures of the Pe-Mn-3 master alloy, oil quenched specimen,

chill casting, and splat are shown in Figure 23, Similarly, the unetched

microstructure of 4330 alloy is shown in Figure 24. As can be seen, the

inclusions in both alloys are refined by increased cooling rate. In both

alloys a tendency toward more spherical inclusion morphology Is observed

with Increased coollig rate. Tn the Fe-Mn-3 alloy the inclusion morphology

changes from wholly dendritic in tie master alloy to wholly spherical in

-he chill casting. Figure 23.

Using the same procedure as was used for the Fe-0.05^31 alloy,

an Po-25^Nl-0,05!«Si alloy was cast in a silica crucible in the vacuum

induction furnace. Reagent grade purity silicon was added to the high

21

purity "zone alloyed" Pe-25^N1 alloy yielding a matrix with clearly defined

solute dendrites and only pure SlOp inclusions. Microstructures of the

master alloy, oil quenched specimen, chill casting, and splat are shown

in Figure 25. In the structure of the Pe-O.05^31 alloy no clustering of

SiOo Inclusions is ever observed. Figure 20, In the structure of the

Pe-eg^Ni-0.05^31 alloy, where the dendrites solidify over a wider temperature

range, the result of solute dendrites is pronounced clustering of SlOp

inclusions. Figure 25. Roughly the same amount of undercooling occurred

in each master alloy, 20 0, as measured by the thermocouple immersed in

the melt. As can be seen, both the matrix and the inclusions are refined

by the higher cooling rates.

22

Discussion

The mlcrostructures investigated herein are necessarily end

products of the various solidification processes. Matrix end inclusion

structures occurring at the early stages of solidification are not

necessarily observed here; only the effect of the final vlctor(s) in

the competition of the structure determining factors has been observed.

The slow cooled master alloys of the Pe-25^Ni, 440G, and 4330 alloys are

all so coarse that the original dendritic microstructure is hardly

recognizable at high magnification. Figures 11, 12, and 13. In the oil

quenched specimens and chill castings the dendritic structures become more

refined and well defined. The splat structures appear as rod-like dendrites

for the Pe-2556N1 and 440C alloys, and equiaxed for the 4330 alloy. The

lamellar eutectic structure of the Pe-4.3^C alloy. Figure 16, Is simply

refined to finer lamellar spacing with higher cooling rates.

The PeO and SiCL inclusions in Figures 19 and 20 are refined

from 3-5 microns diameter in the master alloy to submlcron sizes in the

faster quenched specimens. Qualitatively similar trends are observed for

the inclusions in the Pe-Mn-S and 4330 alloys. Figures 23 and 24. Refinement

of both dendrites and inclusions is observed in Figure 25 for the

Fe-25^Ni".0.05^Si alloy.

The linearity of the plots of log structure parameter (for solute

dendrites, inclusions, and lamellar eutectic) vs. log cooling rate indicates

that the same factor is determining the final solidification structure

over the range of cooling rates observed. As demonstrated by the

isothermal holding in the liquid-solid region experiments of Coughlin,

25 Kattamls, and Plomlngs , this structure determining factor is coarsening

kinetics. The driving force for coarsening kinetics is the reduction

23

25 of surface energy . Thus, the original finely branched dendrite or

small Inclusion that grows into the undercooled liquid eventually

reduces its free energy by reducing its surface area. The result Is

a larger dendrite arm spacing or a larger mean inclusion size.

24

REFERENCES

1. P. Duwez, "Structure and Properties of Alloys Rapidly Quenched From the Idquld State", Trans, of the ASM, v. 60, 1967, pp. 605-633.

2. H. Lou, and P. Duwez, "Face Centered Cubic Ccbalt-Rlch Solid Solutions in Binary Alloys with Aluminum, Gallium, Silicon, Germanium, and Tin" Canadian Journal of Physics, v. 41, 1963, pp. 75^-761.

3. P. Duwez, R. H. Willens, and W. Klement, Jr., "Continuous Series of Metastable Solid Solutions in Silver-Copper Alloys", J. Appl. Phys., v. 31, i960, p. 1136.

4. C. C. Chou, H. L. Lou, and P. Duwez, "CsCl-Type Compounds in Binary Alloys of Rare-Earth Metals with Gold and Silver", Journal of Applied Physics, v. 34, No. 7, 1963, PP. 1971-1973.

3. H. L. Lou and P. Duwez, "Metastable Amorphous Phases in Tellurium Base Alloys", Applied Physics Letters, v. 2, Mo. 1, 1963» P. 21,

6. P. ?redeckl, A. W. Mullendore, and N. J. Grant, "A Study of the Splat Cooling Technique", Trans. Met. Soc. AIME, v. 233» 1965» PP. 1581-1586.

7. R. C. Ruhl, "Cooling Rates In Splat Cooling", Materials Science and Engineering, 1 (1967), pp. 313-320.

8. R. C. Ruhl and M. Cohen, "Splat Quenching of Iron-Carbon Alloys", Trans. Met. Soc. AIME, v. 245, 1969, pp. 241-253.

9. R. W. Strachan, "A Technique for Levltatlon Melting, Undercooling, and Splat Cooling of Metals and Alloys", Ph.D. Thesis, Dept. of Metallurgy, M.I.T., 1967.

10. P, Duwez and S. C. H. Lin, "Amorphous Ferromagnetic Phase in Iron- Carbon- Phosphorous Alloys", U, 3. Atomic Energy Commission Report, 1967.

11. ibid. (9)

12. R, K. Linde, "Kinetics of Transformation of Metastable Silver-Copper Solid Solutions Quenched from the Liquid State", Transactions AIME, v. 236, Jan. 1966, pp. 58-64.

13. H. Matyja, B. C. Glessen, and N. J. Grant, "Effect of Cooling Rate on Dendrite Spacing in Aluminum Alloys", to be published.

14. M. C. Flemings, Solidification Processing class notes.

15. M. C. Flemings: Proceedings Twelfth Sagamore Army Materials Research Conference, August 24 to 27, 1965.

25

16. T. Z. Kattamls and M. C. Flemings: Trans. TOS-AIME, 1965, vol. 233, PP. 992-99.

17. P. C. Quigley and P. J. Ahearn: Trans. An. Foundrymen's Soc,, 1964, vol. 72, pp. 813-17.

18. ibid (15)

19. C. P. Jatczak, D. J. Glradi, and E. S. Rowland: Trans. Am. Soc. Metals, 1956, vol. 48, p. 279.

20. E. Okress, D. Wroughton, G. Comenety, P. Brace, J. Kelly, "Electro- magnetic Levltatlon of Solid and Molten Metals", J. Appl. Fhys., 23, 545-552 (1952).

21. S, Y. Shiriashl and R. 0. Ward, "The Density of Nickel In the Super- heated and Supercooled Liquid States", Can. Met. Quarterly, 3, 117-122 (1964).

22. M. C. Flemings, R. V. Barone, and H. D. Brody, "Investigation of Solidification of High Strength Steel Castings", AfWRC Report, Oct., 1967.

23. S. Bergh and 0. Llndberg, Jemkont. Ann., v. 146, 1962, pp. 862-868.

24. J. W. Cahn and J. E. Billiard, "An Evaluation of Procedures in Quantitative Metallography for Volume Fraction Analysis", Trans. AIME, v. 221, April, 1961.

25. T. Z. Kattamls, J. C. Coughlln, and M. C. Flemings, "Influence of Coarsening on Dendrlte Arm Spacing of Aluminum-Copper Alloys", Trans. Met. Soc. AIME, v. 239» 1967, PP. 1504-1511.

26. G. Forward, "Nucleatlon of Oxide Inclusions During the Solidification of Iron", Sc.D. Thesis, M.I.T., 1966.

IJ. J. Yarwood, "Inclusion Formation in the System Fe-PeO-PeS", Ph.D. Thesis, M.I.T., 1968.

y6

PRISM

GAS OUfLET

TO POWER SOURCE

ENCLOSURE CONTAINING CHARGES AND INGOT MOLDS

GLASSTUBE 5/8"i.<l

LEVITATION COIL

LEVITATED CHARGE

CHARGE CONTAINER

INGOT MOLD

TURNTABLE

- GAS INLET

TO VACUUM PUMPS

HELICAL DRIVING COIL

tO SPLAT COOLER POWER SOURCE

ALUMINUM DRIVING DISC

SPLAT COOLER CoppER pLATE NS

Figure 1-1. Sketch of levltation melting and casting apparatus.

27

Figure 1-2. Photograph of levitation melting and casting apparatus.

28

Figure 1-3. Photograph of levltation coll.

S9

LEPEL INDUCTION GENERATOR

10 KW

LEVITATION COIL

^

COAXIAL LEAD

CAPACITOR BANK

Figure 1-4. Schematic diagram of levitation melter circuit.

30

TWO COLOR PYROMETER

PRISM

1FLAT GLASS DISC

LEVITATED CHARGE

LEVITATION COIL

Figure 1-5. Schematic diagram of temperature measuring system.

31

i, ,J

GAS; EXIT- e

VACUUM RELEASE

VALVE

STANDARD VACUUM - FITTING

> -LEVITATION COIL

COOLING GAS

ENCLOSURE CONTAINING CHARGES AND INGOT MOLDS

THERMOCOUPLE VACUUM

GUAGE

STANDARD VACUUM FITTING

BLEED VALVE

GAS FLOW RATE VALVE AND METERS

COPPER TUBING

MECHANICAL PUMP

\ VACUUM ) VALVE

GATE VALVE

DIFFUSION PUMP

Pigu:?e 1-6. Scheinatlc diagram of gas flow system.

32

Figure 1-7. Photograph of "hammer and anvil",

33

♦'!. \y^p~

Figure 1-8. (a) Optical photocell eye cooling curve, eye in the moving platen, Iron-25 per cent Nickel specimens, clapper velocity 450 cm/second, vertical scale 0.1 volt/division, 1 millisecond/division time scale, 140-160 C superheat,

(b) Resulting microstructure of Pe-25^Ni alloy splat. Magnification 1000X. (Marble's etch).

34

Figure 1-9. Thermocouple cooling curve for Pe~25#Nl alloy 0.08 inches thick chill casting. Dropped from levitation coil at l6000C, 5 millivolt/division vertical scale, 0.5 second/ division horizontal scale.

35

Figure 1-10. Fe-25^Ni chill plate casting 0.08 inches thick, dropped at l600OC from the levitation coil. Magnification (a) 55X, (b) 500X. (Marble's etch).

36

»

iHUfVliil <

• . * » ' V- ■ Alt'"

Figure 1-11. The variation of mlcrostructure with cooling rate for Fe-25^N1 alloy, (a) electron beam "zone alloyed", (b) liquid quenched, (c) chill cast, (d) splat cooled. (Magnification 500X, Marble's etch).

37

(a) (b)

lf> *■ '*

mp:^ ■v;-"'v

lä&B&.f*' *^*' ' - •• ' ^ Wc^J.' Bö ■ • ■ »• ' T' ,'"-* .■**•■'<• - *

mt\iM '*^: - . *■ *

^fe 1 ',.■*' » '. v»

»i^^;.: wV »

RSic i "^ *'"* # W^'Mt^-fi^ , *",/ -. *

(c) (d)

Figure 1-12. Variation of microstructure with cooling rate for 440C alloy, (a) gas quench, (b) liquid quench, (c) chill cast,

(d) splat cooled. (Magnification 500X, Murakami's etch).

38

IT '*^-* 1^* ■ _ *•* ■

(a) (b)

(c) (d)

Figure 1-13« Variation of microstructure with cooling rate for 4330 alloy, (a) gas quench, (b) liquid quench, (c) chill cast,

(d) splat cooled. (Magnification 500X, Rosenhain's etch).

':>

-«b

0) -p 0) u

% lOÄ •H

C^ ^ r-t o o 0) o w o \ o (0

o 3 — (0

u • w 0) >. M > o < «H K g^

■w Ü 1-4 ♦ 2 O TH <) M «l 2

J a o (0 *J O c

~r\

o

Den

drl

te

arm

F

e-25 p

er

ce

u 3 60 •H

(SN0HDJWJ CNIOWS WUV HilHONaü AHVONOOHS

HO

o M r-t

8

f o o ü

10 3 n

$ w c

I 10

in

l

3

O Q - O (L'NüHOIW) oNiovds WHV aiiHONaa ÄUvaNOoas

41

d

Figure 1-16. Variation of microstructure with cooling rate for Pe-4.3^C alloy, (a) master alloy, (b) liquid quenched, (c) chill casting, fd) splat. Magnification 500X. (Nital etch).

42

^

i

a)

U

o o Ü

w

I in

¥

0) M 3 00

(SNOuoiw) ONiovds Hvnawvi Q o

43

«a»«

Figure l-l8. Micrestructure of Pe-0.05^Si master alloy, (a) white light, (b) polarized light. (Magnification 88OX).

44

if

*

>:

Figure 1-19. Microstrueture of Fe-0.07^0 alloy, (a) master alloy, (b) liquid quenched, (c) chill casting, (d) splat. (Magnification 1000X).

45

;■■ ■■ '.'■■■*■. :';;;,

rfi * -SA ■' ..lÄ?i

• •>■■ ■;i<«-. n?-ir

^

/b

* 5 O

Figure 1-20. Variation of microstructure with cooling rate for Pe-0.05#3i alloy. Magnification (a) master alloy, 530X, light microscope, (h) liquid quenched, 1000X, mechanically polished, light microscope, (c) chill casting, 11,200X, electropollshed, parlodlon replica, (d) splat, 1C)0,000X, electropolished, parlodion replica.

46

d = 0.099 MICRONS

CHILL CASTING

d = 0.53 MICRONS

OIL QUENCH

1.0 2.0

fa)

(b)

d = 2.8 MICRONS

MASTER ALLOY

1.0 2.0 3.0 4.0 5.0 6.0 7.0

INCLUSION MEASURED DIAMETER (MICRONS)

Figure 1-21. Size distributions of S10p inclusions in the as cast

Pe-0.05%S1 alloy for the various quenching techniques,

a. chill casting, b. liquid quench, c. master alloy.

47

0) ■p a) U

O O o m 3 W t- (1) >

•p

c o

o C\J rH

O r-<

•H

<U >&

3 « ft) ai E

C cä 0)

CM

I

0)

3 ÖC

•H

o

i a

(SNOUOIW) aaiawvia NOISITIONI NVSW

L

48

/ :

- 4"

% s

k - "I I • * ****** - • '

i

v. \%

1**** ' »..<

. i*V

«.•>

»■ ,*

. > r

v. *. V

d

Figure 1-23. Variation of microstructure with cooling rate for Fe-Mn-S alloy, (a) master alloy (b) liquid quench, (c) chill cast, (d) splat. (Magnification 500X).

49

/

f a

Figure 1-24. Variation of inclusion morphology with cooling rate for 4330 low alloy steel, (a) master alloy, (b) liquid quench, (d) chill cast, (d) splat. (Magnification 500X, unetched).

50

♦■ •.

Figure 1-25. Variation of microstructure with cooling rate for Pe-25^Ni-0.05^Si alloy, (a) master alloy, (b) liquid quenched, (c) chill cast, (d) splat cooled. (Magnification 500X, Marble's etch).

51

Chapter II

ISOTHERMAL SOLID STATE COARSENING OP SILICA INCLUSIONS


After precipitation of a second phase In a solution Is complete,

precipitate size continues to increase due to a phenomenon known as

Ostwald ripening. I.e., reduction of surface free energy by growth of

large particles at the expense of smaller ones. Wagner has developed

a theoretical expression describing Ostwald ripening; a cubic relationship

2 between particle size and ripening time resulted. Greenwood has studied

the ripening of solid spherical precipitates in a saturated liquid solution;

he found a cubic relationship between spherical precipitate radius and

ripening time. Ardell and Nicholson^ in studying the aging kinetics of

gamma prime precipitates in N1-6.71Â1 alloy also found a cubic relationship

between particle size and coarsening (ripening) time. Lifshitz and Slyozov

have applied coarsening kinetics to second phase grain growth in a solid

5 solution. Orlani studied this type of coarsening for spherical carbide

precipitates in iron; again a cubic coarsening law was observed.

High temperature (13710C) homogenization treatments were carried

out on high strength 4330 alloy steel for 24, 64, and 186 hours by Quigley

6 and Ahearn to determine the effect of such treatments on mechanical

properties. An increase of ductility with homogenization time was observed.

Exactly what structural features were altered by this treatment (e.g.,

microsegregation, inclusions, grain size) was not determined, and this work

was, therefore, undertaken to determine if Inclusion ripening (FeO or S109

Inclusions) would be significant at the times and temperatures Involved.

52

nie high temperature solid state coarsening kinetics for SIO«

Inclusions were studied for the Pe-0.055^31 master alloy, chill casting,

anl splat at l400oC. This knowledge will compliment that already obtained

In Section I concerning the solidification behavior of these SlOp ihcluslons

lr- the PB-0. 05^81 alloy.

Since the 310- Inclusions are spherical even at diameters less

than 0,1 P, a simple case of Ostwald ripening of spheres in a solid is

presented as soon as the equilibrium volume fraction of SIO« is reached in

the Fte-0.055*31 al]oy at l400oC.

53

Procedure

Isothermal heat treatment of specimens of the Fe-0.05$3i

alloy at 1400 C + 5 C was carried out In a globar tyyt tube furnace.

The longitudinal temperature gradient in the furnace was first determined,

and the specimens were placed such that the maximum positional temperature

variation was less than + 4 C. Specimens were vacuum encapsulated in

fused quartz tubes after evacuation of the tubes to less than 5 torr and

backfilling with He three times. The quartz tubes were placed in a

firebrick rack which not only enabled removal of the specimens from the

furnace without damage, but also acted as a thermal mass and decreased

both the temporal and positional thermal fluctuations. 1400 C is slightly

below the softening point, but well below the fusing point of fured

quartz, so that, although the evacuated quartz tubes collapsed around the

specimens, the vacuum was maintained. Any leaks in the capsule would

result in total destruction of the specimens at the times and temperatures

used. Such a result was experienced a few times. To prevent a diffusion

reaction between the SiO- tube and the Fe-O.05^31 specimens, each specimen

was wrapped in Pe-0.05$6Si alloy splat. Also, although scmie dirfüslon

bor. :..ig of the splat wrappings and the specimens took place, no bonding i

between the quartz and the splat wrappings was ever observed.

54

Results and Discussion

Specimens of the Fe-O.05^31 master alloy, chill casting, and

splat were held at l400oC for 12, 24, and 48 hours. The resulting size

distributions(Including the "as cast" material for the master and chill

casting) of S10 Inclusions as measured on a polished section are shown

in Figures 1, 2, and 3* These planar distributions may be converted, if

desired, to the true three dimensional distribution as shown in Appendix C

and Figure C-l. For this work it Is assumed that the true mean diameter

of tphei'lcal Inclusions is 1.2X, as discussed in Section I, greater than

the measured planar mean diameter. The volume fraction of inclusions was

again determined by a systematic point count.

Plots of mean inclusion diameter, number of inclusions per unit

area, and volume fraction of inclusions versus time at 1400 C are given in

Figure 4 for the master alloy, chill casting, and splat of the Fe-0.05#Si

alloy. The results of the measurements of the size distributions, volume

fractions, and number per unit area of inclusions (Figures 1, 2, and 3*

and Table 1) indicate that the SlOp inclusions formed in the Pe-0.05^31

alloy coarsen in the solid matrix at 1400 C. The decrease in the number

per unit area (proportional to number per unit volume) and the Increase

in mean S10? inclusion diameter, while maintaining the volume fraction

constant, indicate that the larger inclusions are growing at the expense

of the smaller ones, i.e., Ostwald ripening is taking place. The Increase

of measured volume fraction of SlOp inclusions is the structure during the

first 12 hours of heat treatment indicates either that Qllleon and oxygen

are somewhat supersaturated during solidification and/or subsequent cooling,

or that a small volume fraction of SlOp particles present lie below the optical

55

limit of resolution in the as cast structure. Also, the Increase in volume

fraction of SiO» inclusions after 12 hours may be due in part to an effect

of the increased mean inclusion diameter (radius of curvature) on the

equilibrium solubility of SiOp.

The slightly higher coarsening rate of the master alloy as compared

to that of the chill casting and splat is due to the larger standard deviation

in the as cast size distribution of the inclusions in the master alloy.

Figure la, as compared to that of the chill casting. Figure 2a. This deviation

in particle size is the driving force for Ostwald ripening. A high standard

deviation, therefore, gives a high rate of inclusion coarsening or Ostwald

ripening.

The plots in Figure 4 indicate that appreciable ripening of SiO-

inclusions does occur after 12, 24, and 48 hours in iron at l400oC, approxi-

mately the temperature of Quigley and Ahearn's high temperature homogenization

experiments (1371 C), Since Quigley and Ahearn homogenized for up to 186

hours, Si02 and other types of inclusions in the 4330 steel would have

sufficient time to coarsen appreciably.

56

REFERENCES

1. Wagner, Z. Elchem., v. 65« 196l, p. 58l.

2. G. N. Greenwood, "The Growth of Dispersed Particles In Solutions", Acta. Met., Vol. 4, 1956, pp. 243-248.

3. A. J. Ardell and R. B. Nicholson, "On the Modulated Structure of Aged Nl-Al Alloys", Acta Met., v. 14, 1966, p. 1295.

4. I. M. Llfshltz and V. V. Slyozov, "Kinetics of Diffuse Decomposition of Supersaturated Solid Solution", J. Experimental Theoretical Physics (USSR), Vol. 35, 1958, pp. 479-492.

5. R. Oriani, Acta Met., v. 12, 1964, p. 1399.

6. P. C. Quigley and P. J. Ahearn,"Homogenizatlon of High Strength Steel at 2500 P.", Transactions of the American Btoundrymen's Society.

W7

d = 2.8 MICRONS

60-

3.6 72

I d = 3.7 MICRONS

1—r 10.8 14.4

40

20

0 2.8 5.6

d = 4.1 MICRONS

2.8 5.6 84 11.2 14.0 16.8

INCLUSION MEASURED DIAMETER (MICRONS)

19.6

Figure 2-1. Size distributions of SiOp inclusions in the Pe-0.T#51

master alloy for various homogenization times at H0CoC

100

50

d = 0.1 MICRONS

AS CAST

0 0.4

58

40 -

o M CO

o

100 r

40 r

d = 0.65 MICRONS

0.8 { 1.6 _!_<*= 1.04 MICRONS

0.8 16 2.4

d = 0.89 MICRONS

3.2

0.8 1.6 2.4 3.2 INCLUSION MEASURED DIAMETER (MICRONS)

Figure 2-2. Size distributions of SiOg inclusions In the Pe-O.OS^Sl

chill castings for various homogenizatlon times at l400oC.

59

d = 0.75 MICRONS

«

< Q

2 O M

o

.0 2.4 2.8 32

TNCLÜST0N MEASURED DIAMETER (MICRONS)

60

^40

H 20 to

S 0

MASTER ALLOY

CHILL CASTING AND SPLAT

§ 0.2

0.!

MASTER ALLOY

24 48

10 7r-

6 | 10

EH

g

g icfh

Kf

CHILL CASTING AND SPLAT

O MASTER ALLOY

O CHILL CASTING

A SPLAT

A a.

MASTER ALLOY

.2: ^: ^ Ä

24 HOMOGENTZATION TTME (HOURS)

=a_ 48

Figure 2-4. Variation of Inclusion size fTop). volume fraction Figure Sle)> ^ number per unit area 'Bottom) versus

homogenizatlon time at 1^00 C.

61

Chapter III

MECHANICAL BEHAVIOR OP INCLUSION BEARING IRON


Many studies of the effect of non-metallic Inclusions formed

during solidification on the mechanical properties and fracture behavior of

Iron have been completed. It has been observed that these mlcrocracks

associated with Inclusions are the cause of premature fracture and decreased

ductility In Inclusion bearing alloys, Singh has published a thorough

review of the mechanisms of initiation, growth, and coalescence of mlcrocracks

resulting from either inclusion-matrix interface separation or inclusion

2 fracture during mechanical working. Tipnis and Cook have described deformation

during necking of a tensile bar by means of rigid body rotation, leading in

the case of inclusions to inclusion-matrix interface separation. Liu has

studied Initiation of mlcrocracks by inclusions in spherodized carbon steels.

Of special interest to this work is the prediction by Backofen that cracks

associated with inclusions are formed during heating or cooling due to

differential thermal expansion .

5 Ebeling and Ashby have Investigated the effect of 810- inclusions

on the mechanical behavior of single crystals of copper; they find SlOp

inclusions undeformed after severe matrix deformation. Although the volume

fractions and sizes of inclusions in Ebeling and Ashby's copper alloys were

such that the copper was dispersion hardened, they did observe nn Improvement

of properties with decreasing SiO- inclusion size. Pickering has made a

study of the effects of mechanical working on the deformation of non-metallic

inclusions In the Fe-Si-0 alloy system. Extensively deformed Iron oxide and

62

ferro-slllcate Inclusions were observed optically after rolling. The

forces Induced during deformation at the inclusion-matrix Interface were

concluded to be of prime importance for the deformation and fracture of

7 the Pe-Si-0 alloy. Gell and Leverant studied the effect of MC-type1

carbides on the fatigue properties of Mar-M200 nickel-base superalloy.

They utilized scanning and transmission electron microscopy to examine

fracture surfaces.

Non-metallic inclusions in metal matrices have been observed a

extensively by optical and electron microscopy. 0. Johari has observed

extremely small carbide precipitates in steel using thin film transmission

a electron microscopy. Mlnkoff and Nixon have utilized the great depth of

focus of the scanning electron microscope to observe the morphology of

graphite growth in iron and nickel alloys. Observations of inclusions

in the dimpled region of a ductile fracture surface have been made on the

scanning electron microscope by Baker and Smith in the copper-cuprous

oxide system, and cited by Finniston in his 1968 Hatfleld Memorial Lecture .

In this work, fracture behavior of high purity iron was studied,

as influenced by simple SiO« and FeO inclusions. The volume fractions of

inclusions in the FeO bearing iron was exactly that in the SiO_ bearing

alloy. Table 1. Since both types of inclusions appeared spherical in the

scanning electron microscope at greater than 10,000X, Figures 2 and 5,

the Inclusion shape factor was constant. Also, both types of inclusions

appeared to be randomly positioned in the iron matrix, so this factor was

also constant. 'Ihe two parameters to be studied in this work were inclusion

size and inclucion type (chemistry and structure). Master alloys, chill

ESH

63

castings, and splats with successively finer distributions of inclusions

were tested in tension. Glassy SIO» inclusions and crystalline FeO

inclusions were compared for their effect on mechanical behavior of iron-

64

Procedure

The fracture surfaces of tensile test bars were examined

directly In the scanning electron microscope. Electropollshed and

mechanically polished surfaces of the as cast alloys and the surfaces

of the reduced sections of the test bars were also examined In the

scanning electron microscope. Fractured and electropollshed specimens

.4 were stored In a vacuum of 10 torr before examination on the scanning

electron microscope to prevent any surface contamination. Other than

cutting the specimen to the proper size« no specimen preparation Is

required for the scanning microscope.

65

Results

The first fracture surface to be examined was that of the

base material, Perrovac "E" iron, shown in Figure 1 . As can be seen,

necking continues to a virtual line where the characteristic dimples

of a ductile fracture are evident. Extensive sliding off is observed

in the necked region adjacent to the fracture surface. Even at 6500x,

Figure Id, no inclusions could be observed in the fracture surface

(especially in the dimples) of Perrovac "E".

Next a fractured test bar made from the Pe-O.05^81 master alloy,

mean true Si02 diameter = S.^v was examined. Figure 2 . In this case,

although the macroscopic deformation structure resembles the Perrovac "E"

test bar, many of the dimples on the fracture surface contain spherical

SlOp inclusions. The inclusions are severely displaced, indicating

inclusion-matrix interface separation. No SiO_ inclusions were observed

to be fractured.

Similar behavior on a much smaller scale was observed in the

Fe-0.05»56Si alloy chill casting. Figure 3 . Here the SiO inclusions,

mean true diameter = 0,12tJ , are also perfectly spherical and also are

present in the dimples of the fracture surface. No inclusions could be

observed in the fracture surface of the Pe-0.055631 alloy splat. Figure 4.

The fracture surface of a test bar of the Fe-0,07^0 master

alloy containing Epherlcal FeO inclusions, mean true diameter « 5.2

was obeerved In the rcannln;: electron microscope. Acain, Figure 5,

66

the fracture surface shows dimples containing spherical Inclusions, In

this case FeO. The slightly larger size of PeO inclusions in the

Pe-O.07^0 alloy as compared to the SlOg inclusions in the Fe-0.05£S1

alloy results In the more pronounced void coalescence observed in

Figure 3, as compared to Figure 2 . Similar behavior is evident in the

Fe-0,07%0 chill casting, shown in Figure 6 .. No inclusions could be

observed in the fracture surface of the Pe-0 OTJto splat. Figure 7.

Observations were also made on the necked regions of the fracture

2 surfaces of the master alloys. According to Tipnis and Cook rotation

occurs in the noucentral fracture region, i.e. the shear zone adjacent to

the dimples in the central fracture region. The results of such rotation

are observed in Figures 8,9 for the Fe-C^05&>i and Fe-0.075fo alloys.

That the observed inclusion matrix interface separation has resulted from

rotational type deformation seems apparent. The first impression of "balls

rolling down a trough" is not far from correct, except that the troughs

(voids) are bslng created by the deformation procefls.

Electrope11shed and mechanically polished sections of both the

Fe-0.05%Si and Fe-0.07^0 master alloys were investigated in the scanning

roler*. scope to determine the as cast morphology of the SIO^ and PeO inclusions,

the irjn matrix, and the inclusion-matrix interface. The Pe-O.055^31 master

alloy exhibited intact spherical inclusions with no inclusion-natrix

interface separation. Figure 10, The Fe-0.07^0 master alloy, however,

exhibited extensive inclusion breakage sind Inclusion-matrix separation.

Figure 11. Some of the inclusions also appear to be ferrosillcates instead

of FeO, especially those broken upon .lolldlfication. What are apparently

colidiflcatlon shrinkage holes are observed on the surface of the FeO inclusions,

67

The «Iflctro-polished sections of test bars of the master alloys

and chill castings of the Pe-O.05^31 and Fe-0.07^0 alloys were examined

In the scanning electron microscope. Test bars strained to 5$ elongation

and to necking (then unloaded) as well as a fractured bar of Pe-0.05^31

master alloy exhibited increasing amounts of inclusion-matrix interface

separation with increasing strain in their reduced sections. Figures 12,

13, and 14. No SiO inclusions were observed to be broken. A fractured

test bar of Pe-0.07^0 master alloy exhibited in its electro-polished

reduced section severe inclusion-matrix separation and some severely

fractured PeO Inclusions, Figure 15. This behavior continued well back

of the fracture surface. Behaviors similar to those of the master alloys,

but on a much smaller scale, were observed in the Pe-0.05^81 and Pe-0.07^0

chill castings. Figures 16 and 17.

68

DigQussion

The fracture surfaces of the two alloys studied herein appear very

similar, with slightly more evidence ox void coalescence associated with

the PeO Inclusions. Rotation during the deformation, as observed by

Tipnis and Cook , results in the "ball in a trough" appearance of the

Inclusions and dimples on the fracture surface. The polished reduced sections

of the two alloy test bars did show a marked difference in behavior of PeO

and SiO^. Although inclusion-matrix interface separation was observed in

both ccses, the FeO inclusions were occasionally fractured, while the SiO«

Inclusions were always intact. Figures 14 and 15. This result along with

the resulting microcracks is shown schematically in Figure 18. The work of

Singh in the fracture of aluminum has demonstrated that the lenticular-

shaped void resulting from an Inclusion-matrix Interface separation is easier

to propagate than the blunted void resulting from a fractured inclusion.

7 However, Gell and Leverar.t', working on MC-type carbides in the nickel-base

superalloy, Mar-M200, found that in fatigue testing fractured carbide

incluoions in the as cast Mar-M200 structure were more important in the

initiation of cracks than ralcropores. While the S102 Inclusion bearing iron

showed no voids associated with inclusions in the as cast structure in the

scanning electron microscope, the FeO inclusion bearing iron showed both

inclusion fracture and inclusion-matrix interface separation in the as cast

structure. Figures 10 and 11. Thus the Fe-0.07560 alloy has "cast in" voids

that could cause poor mechanical performance.

Differences in the solidification shrinkage and thermal

expansion of Pe, PeO, and SlOp cause the differing as cast soundness of

the Pe-0.05^Si and Pe-0.075^0 alloys. Since the SlOp inclusions are observed

to be glassy, no abrupt solidification shrinkage and little thermal contrac-

tion occurs. For pure Iron the solidification shrinkage Is several percent,

69

and for PeO the shrinkage Is expected to be larger. The PeO, furthermore,

solidifies well after it is surrounded by solid iron and the solidification

shrinkage should, therefore, appear as a void in the final structure; the

internal voids in Figures 71, 13, and 17 presumably result from this

shrinkage.

Differential thermal contraction between the crystalline or

glassy inclusion particles and the matrix must result in either internal

stresses, inclusion-matrix interface separation, or cracking of the

inclusion or matrix. Data for the thermal expansion of Fe, PeO, and SiO»

are presented in Figure 19* Since high temperature data is not available,

consider thermal expansion due to heating ( L/L = 0 at T = 0oC) of PeO

and SIC in an iron matrix. Assume tue inclusions are inserted into the

matrix at 0 0 with no stresses or voids associated with them. Upon heating

the SlOp inclusions would always be in tension. The PeO inclusions would

expand the same amount as the matrix until the alpha to gamma transformation

where it would go into compression. Just the reverse of the above

hypothetical process should take place upon solidification. SiO is always

in compression upon cooling, while PeO goes into tension at the delta to

gamma transformation and possibly into compression at the gamma to alpha

transformation. This tension could cause the observed inclusion-matrix

interface separation in FeO such as that in Figure 11, or inclusion

fracture in FeO as in Figure 11. Inclusion fracture or inclusion-matrix

interface separation has not been observed in this work for the case of

SiOp inclusions in the as cast Fe-0.05^31 alloy. Figure 10,

Chill casting test bar fracture surfaces. Figures 3 and 6,

resemble master alloy fracture surfaces. Figures 2 and 5» on a much smaller

70

scale for both alloys, i.e., smallei dimples and inclusions In the chill

casting. Although still finer dimples appeared in the splat tensile bar

fracture surfaces. Figures 4 and 7* these are not comparable to the

dimples in the master alloy and chill casting, since no inclusions were

resolvable in the splat fracture surfaces at l4,000X.

TtvB results of the mechanical properties measurements on the

?errovac "E", Fe-0.05$Si and Fe-0.07%0 master alloys given in Appendix D

suggest that the cast-in microcracks associated with the PeO inclusions

in the Pe-0.075^0 alloy as cast structure reduce its luctility as compared

to the relatively high ductility observed in the PB-ü.05#Si alloy, where

no microcracks are associated with the inclusions in the as cast structures.

Additional testing would be necessary to verify this tentative observation.

71

REFERENCES

1. S. N. Singh, "Effects of Ingot Structure and Thermomechanical Processing on Properties of a Wrought High Strength Al-Alloy", Sc.D. Thesis, M.I.T., 1968.

2. V. A. Tlpnls and N. H. Cook, "The Influence of Stress-State and Inclusion Content on Ductile Fracture With Rotation", Journal of Basic Engineering, Sept., 1967, pp. 533-5^0.

3. C, T, Liu and J, Ourland, "The Fracture Beha 'or of Spheroldized Carbon Steel", Trans. ASM, v. 6lf 1968, pp. 156-167.

4. W. A. Backofen, "Fracture of Engineering Materials", ASM, Metals Park, Ohio, 1964, p. 107.

5. M. P. Ashby, "Work Hardening of Dispersion — Hardened Crystals", Phil. Mag. 14, 1966, pp. 1157-1178.

6. P. B. Pickering, "Some Effects of Mechanical Working on the Deformation of Non-metallic Inclusions", Journal of the Iron and Ste»\ Institute, June, 1958, po. 148-159.

7. M. Gell and G. R. Leverant, "The Fatigue of the Nickel-Base Superalloy, Mar-M200, in Single-Crystal and Columnar-Gralnec* Forms at Room Temperature", Trans. Met. Soc. AIME, v. 242, 1968, pp. 1869- 1879.

8. 0. Johari and G. Thomas, "Structures and Strengths of Ausformed Steels", ASM Trans. Quart., v. 58, 1965, pp. 567-578.

9. I. Minkoff and W. C. Nixon, "Scanning Electron Microscopy of Graphite Growth in Iron and Nickel Alloys", J. of Appl. Physics, v. 37, 1966, pp. 4818-4855.

10. C. Baker and G. C. Smith, "Some Observations on the Ductile Fracture of Polycrystalline Copper Containing Inclusions", Trans. Met. Soc. AIME, v. 242, 1968, pp. 1989-1995.

11. H. M. Pinnlston, "Twenty-Five Years On", 'i9th Hatfield Memorial Lecture, Journal of the Iron and Steel Institute, Feb., 1969» pp. 145-153.

72

Figure 3-1. Fracture surface of Perrovac "E" iron 1/2 inch gage length test bar. Magnification (a) 260X, (b) 650X, (c) 2600X, (d) 65OOX.

73

Figure 3-2. Fracture surface of Fe-0.05«S1 alloy 1/2 Inch gage

Sfflcatlon";) 650X, (b) 650X. ( = ) 2600X. (d) 6500X.

74

Figure 3-3. Fracture surface of Fe-0.05^Si chill casting 1/4 inch gage length test bar. Magnification (a) l40X, (b) 1,400X, (c) 7,000X. (d) l4,0CX)X.

75

Figure 3.-4. Fracture surface of Fe-0.05^Si splat 1/4 Inch gage length test bar. Magnification (a) 140X, (b) 2,800X, (c) 7,000X, (d) 14,000X.

76

•

)

Figure 3-5. Fracture surface of Fe-0.07#0 master alloy 1/2 inch gage length test bar. Magnification (a) 85X, (b) 850X, (c) 85OX, (d) l.BOOX.

77

Figure 3-6. Fracture surface of Fe-0.075^0 chill casting 1/4 inch gage length test bar. Magnification (a) 36OX, (b) 1,600X, (c) 8,000X, (d) 16.000X.

78

Figure 3-7. Fracture surface of Pe-0.07^0 splat l/4 Inch gage length test bar. Magnification (a) 700X. (b) 2,800X. (c) 7.000X, (d) 14.000X.

79

Figure 3-8. Necked region of the fracture surface of the Pe-O.05^31 master alloy 1/2 inch gage length test bar.

Magnification: a. 2400X, b. 5500X, c. 12,000X.

80

fi

Figure 3-9. Neoked region of the fracture surface of the Fe-0.07^0 master alloy 1/2 inch gage length test bar. Magnification (a) lf800X, (b) 3,600X, (c) 3.600X.

81

Figure 3-10. Electropollshed surface of the Pe-0.05/^31 master alloy, as cast. Magnification (a) 650X, ^5 degree angle, (b) 6500X, 45 degree angle, (c) 65OX, 90 degree angle, (d) 65OOX, 90 degree angle.

82

Figure 3-11. ElectroTOlished and mechanically polished surfaces of the Fe-Ö.07560 master alloy, as cast, (a) and (b) mechanically polished, (c) and (d) electropolished. Magnification (a) 2,600X, (b) 13,000X, (c) 12,000X. (d) 12,000X.

83

Figure 3-12. Electropolished surface of 0.05^S1-Pfe master alloy 1/2 inch gage length test bar, strained to 5^ elongation and unloaded. Horizontal tensile axis. Magnification (a) 12,0O0X, (b) 12,000X, (c) 22,000X.

84

^^-tf-m

^^v^"^^;'::^ v:tî

W%%$%mim>::: •^?

Figure 3-13. Electropolished surface of the Pe-0.05#S1 master alloy 1/2 Inch gage length test bar, strained to neckirg and unloaded. Horizontal tensile axis. Magnification (a) U.OOOX, (b) 12,000X, (c) 24,OOOX, (d) 24,000X.

85

Figure 3-1^. Mechanically polished surface of the Pe-0.05^Si master alloy 1/2 inch gage length test bar, strained to fracture. Horizontal tensile axis. Magnification (a) 1,250X, (b) 2,600Xt (c) 12,500X, (d) 26,000X.

86

# ^

Figure 3-15. Eleetropollshed surface of Pe-O.OTÔ master alloy 1/2 inch gage length test bar, strained to fracture. Horizontal tensile axis. Magnification (a) 2,400X, (b) 5,500X, (c) 5,50OX, (d) 5,500X.

87

.-i<'M

Figure 3-l6. Electropolished surface of Pe-0.05$Si chill casting l/4 Inch gage length test oar, strained to fracture. Horizontal tensile axis. Magnification (a) ll.OOOX, (b) 22,00OX, (c) 55,OOOX# (d) IIO.OOOX.

88

Figure 3-17. Electropolished surface of Pe-0.07^0 chill casting 1/4 inch gage length test bar, strained to fracture. Horizontal tensile axis. Magnification (a) 24,OOOX, (b) 60,000X, (c) 60,000X, (d) 125,O0OX.

89.

*. *

o OJ

w

■a o

o £ 5

10

s

+i

1 £

*- JO

s si 0 K 2

1 .1- n n *>

9) aaj a)

§iii a» -o w

w c o w

38 9

■P

0) r-i

IS

OJ

CO -H E-"

..CO

00

u 00 •rl

00

äO

2.0

§

*4 . ... a i.o

IRON MATRIX

S102 INCUUSIONS

— PeO INCLUSIONS

1600 2400 TEMPERATURE ( C)

Figure 3-19, Thermal expansion of pure Iron, PeO, and SiO .

91

CONCLUSIONS

1. A levltatlon melting, chill casting, and splat cooling device modified

and developed In this Investigation permits casting of small samples

at cooling rates from 1 C/second to 10 C/secom.

2. Various mlcrostructural features of Iron base alloys were studied.

Including dendrltes, lamellar eutectlc, and non-metallic Inclusions.

These features are all refined by Increasing cooling rate.

3. In the Pe-0,05^31 alloy splat, the SIO^ Inclusions are either surpressed

or made unrecognizably small at 100,000X magnification. Optically

recognizable SIO« Inclusions are found In Pe-0.05^S1 splats after heat

treatment at l400oC.

4. The degree of mlcrostructural refinement of the alloys studied Is a

single valued function of cooling rate over the range of cooling rates

studied. For example, SiOp Inclusion diameters In the Fe-0.05^31

alloy varied from 3 microns at 1 C/second to 0.1 microns at 1500 C/second,

5. SiOp Inclusions in Pe-0.05^S1 alloy have no mlcrocracks associated with

them in the as cast structure. FeO inclusions In Pe-0.07560 alloy show,

in the &z cast structure, (a) inclusion-matrix interface separation,

(b) fracture of FeO inclusions, and (c) cavities within the Inclusions.

The voids associated with the PeO Inclusions result from the volume

changes of the inclusion and matrix during and after solidification.

6. Nearly lOOjS reduction in area was obtained In test bars of high purity

iron, Fe-O.OS^Si alloy, and Pe-0.07^0 alloy. A dimpled fracture surface

la always observed.

92

7. In the inclusion bearing alloys, one or saore incl slons are present

In a large fraction of the dimples, and it Is concluded that the

dimples result from Inclusion-matrix Interface separation during

deformation. Dimples Immediately associated with inclusions are

smaller the smaller the Inclusion. The positions of the inclusions

in the dimples of the fracture surface and in the necked region near

the fracture surface indicate a rotational defonaatlor mode during

necking down.

8. The degree of inclusion-matrix interface separation observed in the

polished reduced sections of the test bars of the Fe-O.OS&l master

alloy Increases with increasing strain.

9. SlOp inclusions in the Pe-0.05$3i alloy coarser, measurably in the solid

state after 12, 24, and 48 hours at l400oC due to Ostwald ripening.

93

a ! o IS o

CM » I o

>, I

SI I-

b I o

• o

• o -a-

• o

CO in

V

c 2

2 I O o

ao 0.

fc I § 3 to

I < O I § PO CO

o w i 8 81

ft-

II 8 o

oc1

CO

o o

g

CVI

m o o

2 o

Ö o

o to

I o

w

i «

a 4)

a> o

M Ä

& & CO

1 • O

i *

o

to

1 : l 1 r-l

& k S^

^ M

p ̂

&

I

o £

94

APPENDIX B. COOLING RATES DURING LIQUID QjJENCHINO

Newton*s Law of cooling for & spherical droplet with "h

controlled11 heat transfer Is written:

hA(T-T)«CdV dT/dt (1) o p

«rfiere T « quenching medium temperature o

T a quenching temperature

A « area of the specimen quenched

V m volume of the speclaen quenched

d ■ density of the specimen quenched

C m specific heat of the specimen quenched p

h ■ heat transfer coefficient

dT/dt a cooling rate at temperature T

Using the above expression, a value for the heat transfer

coefficient during quenching Into a liquid may be calculated from the

measured cooling rates of Bigot and Faivre . Using an oscilloscope to

record the output of a thermocouple embedded In a cylindrical nickel

quenching specimen, they measured a maximum cooling rate in the high

temperature range (800° - 600 C) of 50 C/sec, during quenching from

800 C into oil. Since 8000C is above the decomposition temperature of

quenching oil, no difference in the heat transfer coefficient is expected

for quenching from 800 C (Bigot and Paivre) and quenching from 1550 C

(this work).

Using expression 1 for Bigot and Paivre's work results in a

calculated value of h of 0,00885 cal./cm sec. C, For a one gram

spherical charge of iron in the levitation furnace quenched into oil

95

fron 1550OC:

hdA « 0.0025

where d ■ quenching specimen diameter

k « quenching specimen thermal conductivity.

Therefore, the heat transfer in liquid quenching is "h controlled"« since

hdA Is much less than one. Again using the calculated value of the heat

transfer coefficient and expression 1 for oil quenching of a spherical

one gram levltation melted iron charge quenched from 1550oC, a cooling

rate of 140 C/sec. is calculated.

REFERENCE

1. R. Bigot and R, Paivre, "Application de l'osclllographe cathodlque a l'enreglstrement des courbes temperature-temps au cours de la trempe rapide des metaux". Revue de Metallurgie, LIII, No. 2, 1956, pp. 131-138.

96

APPENDIX C

23 Bergh and Lindberg -* have devised a metltod for converting

a aeasured apparent sphere diameter slê distribution on a polished

surface to the true size distribution. The probability that a spnere

of given diameter will appear as a circle of a given diameter on xhe

plane of polish Is calculated for various sphere and circle sizes.

A table Is then constructed of these values over the range of the size

distribution measured. It Is then assumed that the largest circle on

the plane of polish Is very nearly as large as the diameter of the

sphere It represents. The number of spheres of this diameter that will

appear as this and smaller diameter circles is then calculated from the

probability table and subtracted from the measured values for given

circle sizes, giving a corrected number cf the smaller sized circles.

This process is repeated for each smaller- size circle diameter until

the entire size distribution is converted. Probability tables and a

demonstration of the conversion technique are given in Bergh and Llndberg's

article. The true converted size distribution is compared with the

measured size distribution of PeO inclusions in Figure C-l for the

?e-0.075^0 master alloy.

97

801-

c

d =. 5.2 MICRONS

80r d = 4.4 MICRONS

0 1.6 3.2 4.8 6.4 8.0 9.6

INCLUSION DIAMETER (MICRONS)

11.2 12.8

Figure C-l. Size distribution conversion for PeO inclusions in Pe-0.07560 master alloy. Top: Three dimensional converted size distribution. Bottom: Measured planer size distribution.

98

APPENDIX P. MECHANICAL PROPERTIES

Small flat l/2 Inch gage length by approximately 1/8 inch

thick tensile bars were machined from the master alloys in the vacuum

induction furnace. Tensile testing was carried out on an Instron

machine at 0.05 inches/minute strain rate using a 1/2 inch gage length

extensometer. Results of the tension tests are shown in Table D-l.

The base material. Ferrovac "E", was tested in the as received condition.

Comparisons were made with the measured properties of Ferrovac "E"

Iron: 36,300 psi yield strength, 44,300 psl ultimate, and 26.256

elongation for 1/2 inch gage length bars. The results of the tension

tests on the 1/2 inch gaê length test bars of the Pe-0.05^31 alloy,

mean SiOp diameter »3.4 , indicate the presence of S10. inclusions

slightly lowers the strength said the ductility of the iron. Similarly

for the Fe-0.075^0 master alloy, mean PeO diameter =3.4 , the strength

is lowered, and the ductility is reduced. Since reduction In area in

most cases is 100^ (necking to a point), this parameter was not considered

meaningful.

Observations on the scanning electron microscope of the as cast

Inclusion bearing master alloys indicate that mlcrocracks associated with

the PeO inclusions in the Pe-0.07560 alloy may cause poorer ductility as

compared to the Pe-0.03$Si alloy, where no mlcrocracks are associated with i

the SIC- inclusions in the as cast mlcrostructure. Although the results

of the properties measurements on the two alloys are not directly comparable

due to the substantial difference in yield strengths, one would expect

somewhat lower ductility for the PeO inclusion bearing alloy as compared

to the SIC- Inclusion bearing alloy, as was observed.

99

(0 u

o

Sft. 8°

I g e • I

^H 4»

o » » S 1 •

o • £

ß 5 5 ■s

M

2 rH

in

CVI CO

00

CM

o o

o

f-1 I

Q

i

M 0)

+>

a) w Ü

^

1 ^ 00 o ^

8 o on

8 on

%

M

8 8 8 8 8 g g g 1 on OJ st on ^- o VO ON •« * « it « ft ft * § £1 on $ o -51- S ä £

o

fi

0) +» n cd S

OI

«

OT

UNCLA.S HIEB. Security Classification

I""" DOCUMENT CONTROL DATA - R&D 1 | (Sreunly clmtHllcmlion ol till», body of abstract and irufeumf anrMlalio» mux 6e enlaratf wfian Ihr ove .-„ r<-por( i< claisified

1 1 ORir.lWATIM .-. »CTIVITY fCorporale author;

Massachusetts Institute of Technology Caipbrldge, Massachusetts 02139

■

2a HCPORT iECuS T, CLASSIFICATION 11

Unclassified j 2h GROUP 11

li nePOdT TITLB j

SOLIDIFICATION OF IRON BASE ALLOYS AT LARGE DEGREES OF UNDERCOOLING

1 « DESCRIPTIVE NOTES rTypa of raporr and inclu«iva data»; j

j Interim Report January 1968 - January 1969 | 1 S AUTHORfS) TLaat nama. lint nana. initiml) |

1 Brower,; W. E. Jr. and Flemings, M. C. 1

|6 REPORT DATE

1 July 15, 1969 7a TOTAL NO OF PAeF.5

100 7b. NO. OF REFS I

41 1 ISa. CONTRACT OR GRANT NO.

DA-46-68-C-004A jl fa. PROJECT NO.

D/A 1C02'40^328 1 c

j A,C,S Code No. 5025.11.294 1 "

9a. ORISINATOR-t REPORT NUMBERfS; 1

AMMRC CR 69-14/1

»6. OTHER REPORT HO(S> (Any oth»r numbtr» that may be »laiäned \\ thia raport; |

1 10 A V A IL ABILITY/LIMITATION NOTICES 1

1 This document has been approved for public release and sale; its 1 distribution is unlimited.

Ill SUPPLEMENTARY NOTf <2. SPONSORING MILITARY ACTIVITY

U. S. Army Materials and Mechanics Center! Watertown, Massachusetts 02172 |

113 ABSTRACT 1

j A levitation melting and splat cooling apparatus modified to permit chill 1 casting and liquid quenching was utilized to investigate the effect of | j increased cooling rate on solidification structure and mechanical behavior. | j The degree of microstructural refinement of dendrites, inclusions, and lamellar 1 eutectic with increased cooling rate was determined quantitatively and qualita- 1 1 tively over the range of cooling rates available, l0C/second to 106oC/second for 1 j several iron base alloys. Si02 inclusions in Fe-0.05% Si alloy were observed tc 1 I be measurably coarsened after isothermal solid state heat treatment at 1400oC. I 1 The fracture behavior of inclusion bearing pure iron, in one alloy containing 1

Si02 inclusions, and in the other alloy containing FeO Inclusions, was invest!- 1 gated for various size ranges of inclusions by means of the scanning electron j

i microscope. For both inclusion bearing alloys, the ductile fracture surface j revealed dimples containing spherical inclusions. Extreme inclusion-matrix

interface separation giving a "ball in a trough" appearance resulted from a ! rotational deformation mode during necking down of the test bar. Dimples imme-

diately associated with the inclusions were smaller the smaller the inclusion. 1 1 Although the fracture surfaces of the two alloys were similar, microcracks asso- ] 1 elated with FeO inclusions in the Fe-0.07% 0 alloy were observed in the as-cast i j structure, while no microcracks associated with the Si02 inclusions in the 1 1 Fe-0.05% Si alloy were observed in the as-cast structure. 1 I (Authors) 1

FORM I JAN 64 1473 UNCLASSIFIED

Security Classification

UNCLASSIFIED Security Classification

KEY WORDS

U ndercoollng Splat Cooling Inclusions Solidification Fracture Steel

ROUE WT

LINK 8

«OLE WT

LINK C

ROLE WT

INSTRUCTIONS

I. ORIGINATING ACTIVITY: Enter the name and address of the contractor, subcot tractor, grantee, Department of De- fense activity or other organization (corporate author) issuing the report.

2a. REPORT SECUHTY CLASSIFICATION: Enter the overall security classification of the report. Indicate whether "Restricted Data" is included. Marking is to be in accord- ance with appropriate security regulations.

2b. GROUP: Automatic downgrading is specified in DoD Di- rective 5200.10 and Armed Forces Industrial Manual. Enter the group number. Also, when applicable, show that optional markings have been used for Group 3 and Group 4 as authorized.

3. REPORT TITLE: Enter the complete report title in all capital letters. Titles in all cases should be unclassified. If a meaningful title cannot be selected without classification, show title classification in all capitals in parenthesis immediately following the title.

4. DESCRIPTIVE NOTES: If appropriate, enter the type of report, e.g., interim, progress, summary, annual, or final. Give the inclusive dates when a specific reporting period is covered.

5. AUTHOR(S): Enter the name(s) of authoKs) as shown on or in the report. Enter last name, first name, middle initial. If military, show rank and branch of service. The name of the principal author is an absolute minimum requirement.

6. REPORT DATE: Enter the date of the report as day, month, year; or month, year. If more than one date appears on the report, use date of publication.

7«. TOTAL NUMBER OF PAGES: The total page count should follow normal pagination procedures, i.e., enter the number of pages containing information.

7b. NUMBER OF REFERENCES: Enter the total number of references cited in the report.

8«. CONTRACT OR GRANT NUMBER: If appropriate, enter the applicable number of the contract or grant under which the report was written,

8b, 8c, & Bd. PROJECT NUMBER: Enter the appropriate military department identification, such as project number, subproject number, system numbers, task number, etc. 9a. ORIGINATOR'S REPORT NUMBER(S): Enter the offi- cial report number by which the document will be identified and controlled by the originating activity. This number must be unique to this report. 9b. OTHER REPORT NUMBER(S): If the report has been assigned any other report numbers (either by the originator or by the sponsor), also enter this number(s).

10. AVAILABILITY/LIMITATION NOTICES: Enter any lim- itations on further dissemination of the report, other than those imposed by security classification, using standard statements such as:

(1) "Qualified requesters may obtain copies of this report from DDC"

(2) "Foreign announcement and dissemination of this report by DDC is not authorized."

(3) "U. S. Government agencies may obtain copies of this report directly from DDC. Other qualified DDC users shall request through

(4) "U, S. military agencies may obtain copies of this report directly from DDC Other qualified users shall request through

(5) "All distribution of this report is controlled. Qual- ified DDC users shall request through

If the report has been furnished to the Office of Technical Services, Department of Commerce, for sale to the public, indicate this fact and enter the price, if known. 11. SUPPLEMENTARY NOTES: Use for additional explana- tory notes. 12. SPONSORING MILITARY ACTIVITY: Enter the name of the departmental project office or laboratory sponsoring Cpay ing lor) the research and development. Include address. 13. ABSTRACT: Enter an abstract giving a brief and factual summary of the document indicative of the report, even though it may also appear elsewhere in the body of the technical report. If additional space is required, a continuation sheet shall be attached.

It is highly desirable that the abstract of classified re- ports be unclassified. EacK paragraph of the abstract shall end with an indication of the military security classification of the information in the paragraph, represented as (TS), (S), (C). or (U).

There is no litnitation on the length of the abstract. How- ever, the suggested length is from ISO to 225 words. 14. KEY WORDS: Key words are technically meaningful terms or short phrases that characterize a report and may be used as index entries for cataloging the report. Key words must be selected so that no security c'assification is required. Iden- fieis, such as equipment modei designation, trade name, TUII- t.iry project code name. geoRraphic location, may be used as key words but will be followed hv an indication of technical context. The assignment of links, rules, and weights is optional.

UNCLASSIFIED Security Classification

Documents

SOLIDIFICATION OF IRON BASE ALLOYS of iron base alloys at large degrees of undercooling interim report ammrc cr 69-14/1 by w. e. brower, jr. and m. c. flemings forAuthors: Hasse Fredriksson